𝗔𝗜’𝘀 𝗘𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝗮𝗹 𝗙𝗼𝗼𝘁𝗽𝗿𝗶𝗻𝘁

The facts aren't simple.

Alarmist headlines like “ChatGPT is boiling the planet” may trend, but they rarely tell the whole story. At SHE IS AI™, we believe climate justice and tech equity are inseparable, and that means we need to talk honestly about AI’s environmental cost.

Yes, large models require energy and data centers consume water. But like all tools, AI’s impact depends on how it’s built, who benefits, and whether its development centers people, planet or profit.

One of the best posts I've seen written was by Margarida Barreto where she allowed the numbers to speak for themselves.

𝗠𝗮𝗿𝗴𝗮𝗿𝗶𝗱𝗮'𝘀 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻 𝗽𝗼𝘀𝘁: https://lnkd.in/gzeGsXXm

𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 & 𝗗𝗮𝘁𝗮 𝗖𝗲𝗻𝘁𝗲𝗿𝘀 consume roughly 1–1.3% of global electricity. (IEA)
𝗪𝗮𝘁𝗲𝗿 𝘂𝘀𝗲 𝗶𝘀 𝗿𝗶𝘀𝗶𝗻𝗴. One large model consumed 500,000+ liters in months.
The energy source matters. Renewables drastically lower impact.

But how does that compare?

𝗟𝗶𝘃𝗲𝘀𝘁𝗼𝗰𝗸 𝗔𝗴𝗿𝗶𝗰𝘂𝗹𝘁𝘂𝗿𝗲
Produces 12–17% of global GHG emissions (FAO)
Takes 15,000 liters of water to make 1 kg of beef
Drives 70% of agricultural land use and deforestation

𝗔𝘃𝗶𝗮𝘁𝗶𝗼𝗻
Accounts for 2.5% of global CO₂
One long-haul flight = up to 3 tons CO₂ per person

𝗦𝘁𝗲𝗲𝗹 𝗜𝗻𝗱𝘂𝘀𝘁𝗿𝘆
Responsible for 7–9% of global CO₂ emissions
Massive energy required for smelting + refining

So is AI the villain? No. But like every tool, its footprint depends on how we design it, who benefits, and why it exists.

Here’s what SHE IS AI™ calls for

✔️ Carbon transparency for major models
✔️ Open data on energy + water use
✔️ Low-power AI tailored for the Global South
✔️ Indigenous + local leadership in sustainable tech
✔️ Redefining "scale" to mean smart, not just large

AI isn’t inherently destructive. Although left unchecked, it could deepen inequality and ecological harm.

The solution?
A climate-conscious AI solution built on ethics, creativity, and inclusion.
We must build systems that serve both people and worthy of our world.

We don't need less AI. We need better AI.